scholarly journals Learning to sense from events via semantic variational autoencoder

PLoS ONE ◽  
2021 ◽  
Vol 16 (12) ◽  
pp. e0260701
Author(s):  
Marcos Paulo Silva Gôlo ◽  
Rafael Geraldeli Rossi ◽  
Ricardo Marcondes Marcacini
Keyword(s):  
Vector Space ◽  
Decision Function ◽  
Event Analysis ◽  
Temporal Features ◽  
Variational Autoencoder ◽  
Textual Data ◽  
Decision Making Processes ◽  
Spatio Temporal ◽  
Temporal Events ◽  
One Class Classification

In this paper, we introduce the concept of learning to sense, which aims to emulate a complex characteristic of human reasoning: the ability to monitor and understand a set of interdependent events for decision-making processes. Event datasets are composed of textual data and spatio-temporal features that determine where and when a given phenomenon occurred. In learning to sense, related events are mapped closely to each other in a semantic vector space, thereby identifying that they contain similar contextual meaning. However, learning a semantic vector space that satisfies both textual similarities and spatio-temporal constraints is a crucial challenge for event analysis and sensing. This paper investigates a Semantic Variational Autoencoder (SVAE) to fine-tune pre-trained embeddings according to both textual and spatio-temporal events of the class of interest. Experiments involving more than one hundred sensors show that our SVAE outperforms a competitive one-class classification baseline. Moreover, our proposal provides desirable learning requirements to sense scenarios, such as visualization of the sensor decision function and heat maps with the sensor’s geographic impact.

Fire Detection Based on Fractal Analysis and Spatio-Temporal Features

2021 ◽  
Author(s):  
Monir Torabian ◽  
Hossein Pourghassem ◽  
Homayoun Mahdavi-Nasab
Keyword(s):  
Fractal Analysis ◽  
Fire Detection ◽  
Temporal Features ◽  
Spatio Temporal

Kinematic Adaptive Frequency Sampling Combined Spatio Temporal Features for Snow Monitoring in Aerospace Applications

2021 ◽  
pp. 115472
Author(s):  
Parameshwaran Ramalingam ◽  
Lakshminarayanan Gopalakrishnan ◽  
Manikandan Ramachandran ◽  
Rizwan Patan
Keyword(s):  
Aerospace Applications ◽  
Temporal Features ◽  
Frequency Sampling ◽  
Spatio Temporal

P4-225: Analysis of Spatio-Temporal Features of Histopathology in Murine Alzheimer's Disease Models and Human Alzheimer's Disease Samples

2016 ◽  
Vol 12 ◽  
pp. P1115-P1115
Author(s):  
Vera Niederkofler ◽  
Christina Hoeller ◽  
Joerg Neddens ◽  
Ewald Auer ◽  
Heinrich Roemer ◽  
...  
Keyword(s):  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Disease Models ◽  
Temporal Features ◽  
Spatio Temporal

Predicting Human Mobility with Reinforcement-Learning-Based Long-Term Periodicity Modeling

2021 ◽  
Vol 12 (6) ◽  
pp. 1-23
Author(s):  
Shuo Tao ◽  
Jingang Jiang ◽  
Defu Lian ◽  
Kai Zheng ◽  
Enhong Chen
Keyword(s):  
Reinforcement Learning ◽  
Human Mobility ◽  
Recurrent Network ◽  
Mobility Prediction ◽  
Learning Framework ◽  
Temporal Features ◽  
Wide Range ◽  
Spatio Temporal ◽  
Historical Trajectory

Mobility prediction plays an important role in a wide range of location-based applications and services. However, there are three problems in the existing literature: (1) explicit high-order interactions of spatio-temporal features are not systemically modeled; (2) most existing algorithms place attention mechanisms on top of recurrent network, so they can not allow for full parallelism and are inferior to self-attention for capturing long-range dependence; (3) most literature does not make good use of long-term historical information and do not effectively model the long-term periodicity of users. To this end, we propose MoveNet and RLMoveNet. MoveNet is a self-attention-based sequential model, predicting each user’s next destination based on her most recent visits and historical trajectory. MoveNet first introduces a cross-based learning framework for modeling feature interactions. With self-attention on both the most recent visits and historical trajectory, MoveNet can use an attention mechanism to capture the user’s long-term regularity in a more efficient way. Based on MoveNet, to model long-term periodicity more effectively, we add the reinforcement learning layer and named RLMoveNet. RLMoveNet regards the human mobility prediction as a reinforcement learning problem, using the reinforcement learning layer as the regularization part to drive the model to pay attention to the behavior with periodic actions, which can help us make the algorithm more effective. We evaluate both of them with three real-world mobility datasets. MoveNet outperforms the state-of-the-art mobility predictor by around 10% in terms of accuracy, and simultaneously achieves faster convergence and over 4x training speedup. Moreover, RLMoveNet achieves higher prediction accuracy than MoveNet, which proves that modeling periodicity explicitly from the perspective of reinforcement learning is more effective.

scholarly journals Ensemble reconstruction of spatio-temporal extreme low-flow events in France since 1871

2017 ◽  
Vol 21 (6) ◽  
pp. 2923-2951 ◽  
Author(s):  
Laurie Caillouet ◽  
Jean-Philippe Vidal ◽  
Eric Sauquet ◽  
Alexandre Devers ◽  
Benjamin Graff
Keyword(s):  
Climate Change ◽  
Low Flow ◽  
Reference Network ◽  
Large Set ◽  
Historical Events ◽  
Variable Threshold ◽  
Matching Procedure ◽  
Spatio Temporal ◽  
Temporal Events ◽  
Rich Countries

Abstract. The length of streamflow observations is generally limited to the last 50 years even in data-rich countries like France. It therefore offers too small a sample of extreme low-flow events to properly explore the long-term evolution of their characteristics and associated impacts. To overcome this limit, this work first presents a daily 140-year ensemble reconstructed streamflow dataset for a reference network of near-natural catchments in France. This dataset, called SCOPE Hydro (Spatially COherent Probabilistic Extended Hydrological dataset), is based on (1) a probabilistic precipitation, temperature, and reference evapotranspiration downscaling of the Twentieth Century Reanalysis over France, called SCOPE Climate, and (2) continuous hydrological modelling using SCOPE Climate as forcings over the whole period. This work then introduces tools for defining spatio-temporal extreme low-flow events. Extreme low-flow events are first locally defined through the sequent peak algorithm using a novel combination of a fixed threshold and a daily variable threshold. A dedicated spatial matching procedure is then established to identify spatio-temporal events across France. This procedure is furthermore adapted to the SCOPE Hydro 25-member ensemble to characterize in a probabilistic way unrecorded historical events at the national scale. Extreme low-flow events are described and compared in a spatially and temporally homogeneous way over 140 years on a large set of catchments. Results highlight well-known recent events like 1976 or 1989–1990, but also older and relatively forgotten ones like the 1878 and 1893 events. These results contribute to improving our knowledge of historical events and provide a selection of benchmark events for climate change adaptation purposes. Moreover, this study allows for further detailed analyses of the effect of climate variability and anthropogenic climate change on low-flow hydrology at the scale of France.

scholarly journals Identifying VoIP traffic in VPN tunnel via Flow Spatio-Temporal Features

2020 ◽  
Vol 17 (5) ◽  
pp. 4747-4772
Author(s):  
Faiz Ul Islam ◽  
◽  
Guangjie Liu ◽  
Weiwei Liu ◽  
Keyword(s):  
Temporal Features ◽  
Spatio Temporal

scholarly journals Unsupervised video object segmentation: An affinity and edge learning approach.

2021 ◽  
Author(s):  
Sundaram Muthu ◽  
Ruwan Tennakoon ◽  
Reza Hoseinnezhad ◽  
Alireza Bab-Hadiashar
Keyword(s):  
Optic Flow ◽  
Object Segmentation ◽  
Video Object Segmentation ◽  
Video Object ◽  
Post Processing ◽  
Multiple Objects ◽  
Temporal Features ◽  
Image Edge ◽  
Spatio Temporal ◽  
Generalization Errors

<div>This paper presents a new approach to solve unsupervised video object segmentation~(UVOS) problem (called TMNet). The UVOS is still a challenging problem as prior methods suffer from issues like generalization errors to segment multiple objects in unseen test videos (category agnostic), over reliance on inaccurate optic flow, and problem towards capturing fine details at object boundaries. These issues make the UVOS, particularly in presence of multiple objects, an ill-defined problem. Our focus is to constrain the problem and improve the segmentation results by inclusion of multiple available cues such as appearance, motion, image edge, flow edge and tracking information through neural attention. To solve the challenging category agnostic multiple object UVOS, our model is designed to predict neighbourhood affinities for being part of the same object and cluster those to obtain accurate segmentation. To achieve multi cue based neural attention, we designed a Temporal Motion Attention module, as part of our segmentation framework, to learn the spatio-temporal features. To refine and improve the accuracy of object segmentation boundaries, an edge refinement module (using image and optic flow edges) and a geometry based loss function are incorporated. The overall framework is capable of segmenting and finding accurate objects' boundaries without any heuristic post processing. This enables the method to be used for unseen videos. Experimental results on challenging DAVIS16 and multi object DAVIS17 datasets shows that our proposed TMNet performs favourably compared to the state-of-the-art methods without post processing.</div>

Sign in / Sign up

Export Citation Format

Share Document